Exploiting training example parallelism with a batch variant of the ART 2 classification algorithm

Authors:
Philipp Ciechanowicz;Stephan Dlugosz;Herbert Kuchen;Ulrich Müller-Funk
Affiliations:
University of Münster, Münster, Germany;University of Münster, Münster, Germany;University of Münster, Münster, Germany;University of Münster, Münster, Germany
Venue:
PDCN '08 Proceedings of the IASTED International Conference on Parallel and Distributed Computing and Networks
Year:
2008

Citing 4
Cited 0

Art 2-A: an adaptive resonance algorithm for rapid category learning and recognition

Neural Networks
Using and designing massively parallel computers for artificial neural networks

Journal of Parallel and Distributed Computing - Special issue on neural computing on massively parallel processing
Using MPI (2nd ed.): portable parallel programming with the message-passing interface

Using MPI (2nd ed.): portable parallel programming with the message-passing interface
Parallel Implementations of Backpropagation Neural Networks on Transputers: A Study of Training Set Parallelism

Parallel Implementations of Backpropagation Neural Networks on Transputers: A Study of Training Set Parallelism

Quantified Score

Hi-index	0.00

Visualization

Abstract

In this article we develop a batch variant of the ART 2 classification algorithm invented by Carpenter and Grossberg. Our algorithm exploits training example parallelism while leaving the overall design of the ART 2 network unchanged such that a significant reduction of the execution time can be achieved on a multiprocessor system. We present a parallel implementation strategy and analyze it w.r.t. execution time and speedup. As our algorithm naturally benefits from data parallelism, the implementation uses data parallel skeletons of the Muenster skeleton library Muesli. We show that skeletons are an efficient way to write parallel applications compared to a manual MPI implementation.